1
00:00:00,000 --> 00:00:18,879
36C3 preroll music
2
00:00:18,879 --> 00:00:25,802
Herald: Next is Bijan. Bijan. Bijan, I
pronounce this. Pretty persian. Yeah. He's
3
00:00:25,802 --> 00:00:32,730
an attorney, ein Rechtsanwalt it is called in
deutsch, and he works for the Gesellschaft
4
00:00:32,730 --> 00:00:37,971
für Freiheitsrechte in Berlin. If I'm
right. Good. Give them a welcome.
5
00:00:37,971 --> 00:00:42,369
applause, please. It's early in the
morning. We're going to kick back here.
6
00:00:42,369 --> 00:00:46,459
applause
7
00:00:46,459 --> 00:00:51,390
Bijan: Early in the morning, only at the
Congress you can call 12:30 early in the
8
00:00:51,390 --> 00:00:57,460
morning, but it is. And, um, well, if
you've ever sat on a plane and wondered
9
00:00:57,460 --> 00:01:02,350
what the person three rows behind you is
eating, whether they flying alone, whether
10
00:01:02,350 --> 00:01:07,050
they have checked in their luggage or only
hand luggage and what visa they were using
11
00:01:07,050 --> 00:01:12,510
when they were buying their plane ticket,
then you're probably a police officer or
12
00:01:12,510 --> 00:01:16,909
should join the national police of any EU
member state, because that is exactly what
13
00:01:16,909 --> 00:01:21,673
the national polices in Germany and
Austria and other Europeans member states,
14
00:01:21,673 --> 00:01:25,689
Europe, member states of the European
Union, can do. Thanks to the PNR
15
00:01:25,689 --> 00:01:30,479
directive, which is the topic of today's
talk. And we are going to talk and explain
16
00:01:30,479 --> 00:01:35,719
to you what the PNR directive and the laws
transposing it into national law are all
17
00:01:35,719 --> 00:01:40,889
about, why this is problematic and what we
can do and what we are actually doing
18
00:01:40,889 --> 00:01:46,780
against it in order to stop it. And Walter
will start off with a few infos.
19
00:01:46,780 --> 00:01:53,299
Walter: Yeah. Hello. So firstly, I would
like to introduce into Epicenter Works,
20
00:01:53,299 --> 00:01:58,915
because we have already a history on
bringing down data retention laws. So
21
00:01:58,915 --> 00:02:03,959
probably you know us from our fight
against data retention in Europe when we
22
00:02:03,959 --> 00:02:08,429
still were called "AKA Vorrat Österreich".
I am working for Epicenter Works on a
23
00:02:08,429 --> 00:02:14,610
voluntary basis. And I would like to
mention my colleague Angelika Adensamer
24
00:02:14,610 --> 00:02:19,800
who did the main work on this for
Epicenter Works. But she cannot be at
25
00:02:19,800 --> 00:02:31,130
Congress this year. So, flight data. It is
said, I've heard that at any given point
26
00:02:31,130 --> 00:02:38,554
in time, one million people are on a plane
in the skies flying around the globe. As
27
00:02:38,554 --> 00:02:47,379
you can see here. And today, although in
times of resource exhaustion, we should
28
00:02:47,379 --> 00:02:52,959
talk about that anyway. I am convinced
today we are talking about the data
29
00:02:52,959 --> 00:03:00,371
protection issue about it. A big one. And
we are talking about passenger name
30
00:03:00,371 --> 00:03:07,560
records. So what is a passenger name
record, anyway? A passenger name record,
31
00:03:07,560 --> 00:03:19,360
as you can see here, is a data set
compiled of 19 different data fields. So
32
00:03:19,360 --> 00:03:26,000
you can get about up to 60 different data
points on one single passenger on one
33
00:03:26,000 --> 00:03:31,549
single flight. So, for instance, you have
data in there like the first and second
34
00:03:31,549 --> 00:03:40,109
name, address, but also other things,
metadata. More important things, like the
35
00:03:40,109 --> 00:03:47,340
means of payment you made, the point in
time when you booked the flight and things
36
00:03:47,340 --> 00:03:54,280
like that. And as a specific problem about
it is that there is also a free text field
37
00:03:54,280 --> 00:04:02,170
so airline employees can enter data there
and which we cannot control. And
38
00:04:02,170 --> 00:04:11,739
altogether we have a quite big data set of
each passenger on each flight. So this is
39
00:04:11,739 --> 00:04:19,921
common in the airline industry. But in
2016, the PNR directive came about. So
40
00:04:19,921 --> 00:04:26,385
what is the PNR directive? It is a piece
of European legislation , which was
41
00:04:26,385 --> 00:04:35,080
enacted in April 2016. And when we have
European legislation, it's important to
42
00:04:35,080 --> 00:04:41,980
mention that it doesn't come out of the
blue out of Brussels, but it is enacted
43
00:04:41,980 --> 00:04:49,563
together with from the commission, the
European Parliament and the council. And
44
00:04:49,563 --> 00:04:56,070
the council are the governments of our
member states. So we have to keep in mind
45
00:04:56,070 --> 00:05:03,024
that member states governments, have a
big say when things like this are enacted.
46
00:05:03,024 --> 00:05:10,756
And it is a directive. And that means that
every single member state has to transpose
47
00:05:10,756 --> 00:05:16,980
the content of the directive into its
national law. And this had to be done
48
00:05:16,980 --> 00:05:24,824
until the 25th May of 2018. This was the
the tenth transposition deadline. And for
49
00:05:24,824 --> 00:05:32,661
instance, Austria and Germany made laws to
transpose that into their national law. So
50
00:05:32,661 --> 00:05:43,590
what had they to enact? They had to enact
laws prescribing that all airlines have to
51
00:05:43,590 --> 00:05:51,986
transfer data of all passengers, all
passenger name records of every flight,
52
00:05:51,986 --> 00:05:59,100
and they have to be pushed to a national
police database. So unlike the telecom
53
00:05:59,100 --> 00:06:06,290
data retention I already mentioned, the
data is not kept where it where where it
54
00:06:06,290 --> 00:06:11,380
is created. But it has to be pushed from
the private sector, from the airlines, to
55
00:06:11,380 --> 00:06:21,060
police database, databases. And the data
retention directive prescribes that every
56
00:06:21,060 --> 00:06:26,788
flight leaving or entering the European
Union must be covered by that. But in
57
00:06:26,788 --> 00:06:32,649
addition, every single member state also
covered flights within the EU. So you have
58
00:06:32,649 --> 00:06:37,880
we have the full take now. Flights within
the EU as well as flights leaving or
59
00:06:37,880 --> 00:06:46,060
entering the EU. And every single record
of every single passenger of every single
60
00:06:46,060 --> 00:06:55,410
flight is in a police database and will be
compared with existing databases, for
61
00:06:55,410 --> 00:07:02,573
instance, of known criminals or of stolen
passports and the like. And they try to
62
00:07:02,573 --> 00:07:10,560
find matches there. And what they are also
going to do is matching with predetermined
63
00:07:10,560 --> 00:07:16,245
criteria. So they will come up with flight
patterns of known perpetrators, for
64
00:07:16,245 --> 00:07:22,350
instance, when they booked a flight and so
on. They will algorithmically try to find
65
00:07:22,350 --> 00:07:29,550
patterns there, and then they will compare
your flight passenger name records with
66
00:07:29,550 --> 00:07:34,470
that data. And if you have a similar
behavior, than a previous perpetrator,
67
00:07:34,470 --> 00:07:42,348
previous criminal, for instance, then
you're already under suspicion. And this
68
00:07:42,348 --> 00:07:47,880
data in these databases are stored for
five years and can be further used by
69
00:07:47,880 --> 00:07:54,460
different law enforcement agencies. So
that data is not only compared and then
70
00:07:54,460 --> 00:08:00,560
deleted again. The storage time is five
years and they do something called
71
00:08:00,560 --> 00:08:08,040
depersonalization about six months after
the data was created. But this is not in
72
00:08:08,040 --> 00:08:13,480
any way an anonymisation, but they just
remove some data and it can easily be
73
00:08:13,480 --> 00:08:20,530
identified again. So the person the data
belongs to can easily be identified for
74
00:08:20,530 --> 00:08:27,495
the whole period of five years. So you
probably asked yourself already: First, is
75
00:08:27,495 --> 00:08:33,958
this effective? Well, this runs already
since last year, so we have some data.
76
00:08:33,958 --> 00:08:43,714
First, I will present to you the data from
Austria. In Austria, we found out that
77
00:08:43,714 --> 00:08:52,220
already until the 30th of September, 2019,
almost 24 passenger name records where
78
00:08:52,220 --> 00:08:59,450
forwarded to the passenger name unit at
the Bundeskriminalamt and
79
00:08:59,450 --> 00:09:06,640
11 900 000 thousand different people
were subject to that. And of these, almost
80
00:09:06,640 --> 00:09:13,005
24 000 000 passenger name records, the
algorithms that checking against databases
81
00:09:13,005 --> 00:09:21,210
already brought up 190 000
matches. So every single match,
82
00:09:21,210 --> 00:09:27,639
every single output the algorithm has,
must be checked by a human employee. So we
83
00:09:27,639 --> 00:09:34,430
have sitting there people who have to
check. Even this is not even the data of a
84
00:09:34,430 --> 00:09:41,580
year. And they have to check
190 000 matches and only 280
85
00:09:41,580 --> 00:09:47,340
of them are actual hits. So if a person
checks what the algorithm outputs there,
86
00:09:47,340 --> 00:09:55,540
then only in 0.15% of the cases
the policewoman or policeman
87
00:09:55,540 --> 00:10:01,610
come to their conclusion: This is actually
relevant for us. And if you do the math,
88
00:10:01,610 --> 00:10:09,731
this means that only 0.001% percent of all
that 24 million passenger name data, your
89
00:10:09,731 --> 00:10:15,810
data which is checked, actually leads to a
hit. And we don't even know how many
90
00:10:15,810 --> 00:10:23,120
actual false positives remain in these
220. This is only what the police will
91
00:10:23,120 --> 00:10:29,980
inspect afterwards. So we have no numbers
or results if they had actual
92
00:10:29,980 --> 00:10:36,840
investigative results on that. But what we
can say is that there are 21 employees,
93
00:10:36,840 --> 00:10:42,070
qualified employees, working in the
passenger name, Passenger Information Unit,
94
00:10:42,070 --> 00:10:48,880
and this costs almost 2 million euros per
year and only for checking that data in
95
00:10:48,880 --> 00:10:55,050
the small country of Austria. And Bijan
now will present to you the data in German.
96
00:10:55,050 --> 00:10:59,285
Bijan: The number, the data of the big
neighbor, because you said small country
97
00:10:59,285 --> 00:11:05,610
Austria. In Germany the numbers are
surprisingly similar. We also had - have
98
00:11:05,610 --> 00:11:11,341
numbers up until mid of August 2019, and
we have had almost 32 million passenger
99
00:11:11,341 --> 00:11:16,980
name records checked, which generated
automatic results of matches of about
100
00:11:16,980 --> 00:11:24,290
240 000, which then were checked by 40
police officers and there remained only
101
00:11:24,290 --> 00:11:32,910
910 actual hits. So the fail rate was
99.6% and 0.003% all PNRs checked led to
102
00:11:32,910 --> 00:11:38,090
actual hits. And even of that number, just
as in Austria, we are not sure how many
103
00:11:38,090 --> 00:11:43,170
false positives remain. We know that there
were considerably a considerable amount of
104
00:11:43,170 --> 00:11:47,670
false positives. We estimate them to be in
the hundreds. But the law enforcement did
105
00:11:47,670 --> 00:11:52,650
not specify what actually, how many
supposed positives remained, even among
106
00:11:52,650 --> 00:11:59,202
the 910. And one of the results we know is
that it led to 57 arrests. We don't know
107
00:11:59,202 --> 00:12:03,000
for which crimes. We don't know whether
these people actually committed a crime,
108
00:12:03,000 --> 00:12:08,190
whether they were suspected for crime,
whether they were just on a watch list.
109
00:12:08,190 --> 00:12:15,220
But 57 arrests, assuming this is these
were legitimate, this means that 0.0002%
110
00:12:15,220 --> 00:12:21,430
of all PNRs checked led to an arrest. And
if you try to to transpose this to other
111
00:12:21,430 --> 00:12:28,150
situations in life, you could go to a to a
market, to to some, uh, to some festival
112
00:12:28,150 --> 00:12:32,490
or what not, and just ask randomly people,
and you would probably have with a similar
113
00:12:32,490 --> 00:12:36,990
probability, an arrest in the end at the
end of the day. So if this holds that this
114
00:12:36,990 --> 00:12:41,735
whole PNR processing holds is this
effectiveness is the standard that we are
115
00:12:41,735 --> 00:12:49,611
happy with, then you can easily take this
to all other sorts of walks of life. And
116
00:12:49,611 --> 00:12:55,650
this is true, in our opinion, a big
problem, because it will lead to a digital
117
00:12:55,650 --> 00:13:00,766
surveillance state, which is has come
quite near with these new tools that the
118
00:13:00,766 --> 00:13:07,510
PNR directive provide. What we've now just
shown are the the automatic is the checks
119
00:13:07,510 --> 00:13:12,246
against databases. That was the one thing
that the PNR directive provides for. The
120
00:13:12,246 --> 00:13:17,070
other one is the checking against
predetermined criteria. And this is where
121
00:13:17,070 --> 00:13:23,010
the voodoo kind of starts, because the
idea that you can merely from the data
122
00:13:23,010 --> 00:13:29,145
that is in the PNR, in your passenger name
record, derive whether you are suspicious,
123
00:13:29,145 --> 00:13:36,292
or dangerous even is, at least in our
opinion, pretty much voodoo, and it has
124
00:13:36,292 --> 00:13:43,089
serious consequences. And it might lead to
automatic profiling affecting hundreds of
125
00:13:43,089 --> 00:13:47,270
millions of people, possibly, because
everybody is checked when they and when
126
00:13:47,270 --> 00:13:51,830
they use a plane. Everybody PNR record is
checked against these automatic , against
127
00:13:51,830 --> 00:13:56,742
these predetermined criteria, and not just
for crimes such as terrorism or organized
128
00:13:56,742 --> 00:14:01,670
crime, where you could maybe make a case
that there exists such a thing as a
129
00:14:01,670 --> 00:14:06,887
pattern of movements where you can
identify a terrorist suspect, but it is
130
00:14:06,887 --> 00:14:13,850
also used for crimes such as fraud or
forgery or cyber crime where I would argue
131
00:14:13,850 --> 00:14:18,501
you cannot find the typical cyber
criminals flight pattern, flight patterns.
132
00:14:18,501 --> 00:14:24,339
It's just not possible. And so but but the
PNR directive itself is only the one
133
00:14:24,339 --> 00:14:29,120
thing. We are fighting this for reasons
that go way beyond the PNR processing so
134
00:14:29,120 --> 00:14:35,980
the processing of PNR flight data, because
it may set a dangerous precedent for other
135
00:14:35,980 --> 00:14:40,870
mass surveillance. Already now PNR
processing is being discussed for buses
136
00:14:40,870 --> 00:14:44,850
that cross borders, for ships and trains.
And there are some countries such as
137
00:14:44,850 --> 00:14:50,870
Belgium that have already enacted the very
much. And why stop there, might a police
138
00:14:50,870 --> 00:14:56,220
officer argue. Why not include rental cars
that cross borders? Why not at some point
139
00:14:56,220 --> 00:15:00,410
include private cars that cross borders?
Why not get away with that requirement of
140
00:15:00,410 --> 00:15:04,460
crossing borders? Why not have everybody
checked all the time, maybe via their
141
00:15:04,460 --> 00:15:09,899
mobile phones? So when we give way to this
sort of data processing with such a low
142
00:15:09,899 --> 00:15:16,040
threshold of effectiveness, we open the
door for all sorts of, um, of activity
143
00:15:16,040 --> 00:15:21,310
that at least from our point of view, is
illegal. And the question you were maybe
144
00:15:21,310 --> 00:15:27,311
asking yourself or maybe not. Is this
legal? We are convinced it is not. And
145
00:15:27,311 --> 00:15:35,071
luckily, we could rely on a legal opinion
that the European Court of Justice ECJ has
146
00:15:35,071 --> 00:15:40,800
rendered a two and a half years ago. There
is one PNR agreement in place between the
147
00:15:40,800 --> 00:15:46,200
EU and the USA, which has not been
challenged yet. And another agreement was
148
00:15:46,200 --> 00:15:51,330
supposed to be known or was negotiated
between the EU Commission and Canada, and
149
00:15:51,330 --> 00:15:56,600
the EU Parliament then presented the
question to the ECJ whether this agreement
150
00:15:56,600 --> 00:16:01,830
would be violating fundamental rights of
the Charter of Fundamental Rights of the
151
00:16:01,830 --> 00:16:09,360
European Union. And the ECJ concluded that
it would, in the form that it was proposed
152
00:16:09,360 --> 00:16:13,800
to it, breach Article 7 and 8 of that
charter's. Article 7 as the right to
153
00:16:13,800 --> 00:16:17,971
privacy in Article 8 is the right to have
your data protected, your personal data
154
00:16:17,971 --> 00:16:23,746
protected. And we are, of course, relying
heavily on that, on the arguments that the
155
00:16:23,746 --> 00:16:29,149
court developed and developing them even
further, because; as you can imagine, the
156
00:16:29,149 --> 00:16:34,563
PNR, the agreement with Canada and the PNR
directive are quite similar. So what are
157
00:16:34,563 --> 00:16:40,821
these arguments that we are bringing up?
And we've shown already that the
158
00:16:40,821 --> 00:16:45,884
effectiveness is highly doubtful. And this
leads us to concluding that the PNR
159
00:16:45,884 --> 00:16:50,432
directive is disproportionate. So it
violates human fundamental rights. For
160
00:16:50,432 --> 00:16:55,943
several reasons. One being a point that
we've both raised already that PNR
161
00:16:55,943 --> 00:17:00,409
processing indiscriminately affects all
passengers. And this is a very important
162
00:17:00,409 --> 00:17:05,069
point, because it makes it shows the
difference between PNR processing under
163
00:17:05,069 --> 00:17:08,920
the PNR directive and what was formerly
the the data retention of
164
00:17:08,920 --> 00:17:14,720
telecommunications data. Because the
latter would require a specific case,
165
00:17:14,720 --> 00:17:20,029
something must have had happened in order
for the law enforcement to ask for the
166
00:17:20,029 --> 00:17:26,010
telecommunications data of the
telecommunications provider. But our
167
00:17:26,010 --> 00:17:32,210
PNR data on flights is checked all the
time, always, against databases, and even
168
00:17:32,210 --> 00:17:36,669
more importantly, the predetermined
criteria, which we, of course, do not know
169
00:17:36,669 --> 00:17:41,600
nothing about. And this brings with it
especially the last point, the
170
00:17:41,600 --> 00:17:46,205
predetermined criteria, are high risk of
false accusations. We've already seen that
171
00:17:46,205 --> 00:17:52,731
99.6% of data base matching, automatic
data is matching is wrongful. And imagine
172
00:17:52,731 --> 00:17:59,639
how much higher the number would be with
checking against predetermined criteria.
173
00:17:59,639 --> 00:18:05,809
And that the reason why we expect many
false accusations, false positives, is the
174
00:18:05,809 --> 00:18:09,379
so-called base rate fallacy, which
basically says that when you're looking
175
00:18:09,379 --> 00:18:13,980
for a very small amount of people in a
large dataset and you have a significant
176
00:18:13,980 --> 00:18:18,769
fail rate, you're very likely to produce
more false positives, maybe many more
177
00:18:18,769 --> 00:18:23,660
false positives than true positives. So
actual suspects, or not suspects, but
178
00:18:23,660 --> 00:18:28,104
actual terrorists. So, for instance, when
you if you're checking 100 million flight
179
00:18:28,104 --> 00:18:32,700
passengers. And you're looking for 100
terrorists, and you have even a fail rate
180
00:18:32,700 --> 00:18:39,659
of 0.1%, not the 99.6 that we're talking
about now, but even just 0.1%, this would
181
00:18:39,659 --> 00:18:45,269
render this would this would render
100 000 flight passengers subject to to
182
00:18:45,269 --> 00:18:50,756
to being suspected terrorists. So you
would have 100 000 false positives, 100
183
00:18:50,756 --> 00:18:55,280
terrorists that let's assume all of them
so that they had a positive success rate
184
00:18:55,280 --> 00:19:00,311
of 100 percent identifying positively as a
terrorist suspect. Then you will have
185
00:19:00,311 --> 00:19:07,291
100 000 false positives, 100 people that
are correctly suspected. But everybody, of
186
00:19:07,291 --> 00:19:11,399
course, will be treated the same. And what
I've listed here are just the obvious
187
00:19:11,399 --> 00:19:16,529
things, stigmatization at the airport by
interrogation, searches of luggage of
188
00:19:16,529 --> 00:19:21,460
people and arrests, missing flights.
And depending on the country
189
00:19:21,460 --> 00:19:27,019
you're in you may be in much more trouble
after that. The second point is that the
190
00:19:27,019 --> 00:19:33,075
data is being stored way too long. As
Walter has already mentioned 5 years. Why
191
00:19:33,075 --> 00:19:38,549
do you need 5 years worth of data to check
a database entry or against a
192
00:19:38,549 --> 00:19:42,795
predetermined criteria? Of course, you
don't needed it for that. Because you
193
00:19:42,795 --> 00:19:47,970
could do that immediately after a person
has boarded. You can perform the check and
194
00:19:47,970 --> 00:19:52,584
then you could get rid of the data, delete
it after it's being used. The reason why
195
00:19:52,584 --> 00:19:56,508
they're storing it so long as that law
enforcement and intelligence agencies have
196
00:19:56,508 --> 00:20:01,489
an interest that goes beyond that checking
after boarding, they want to keep the data
197
00:20:01,489 --> 00:20:06,649
and check it in future, criminal
investigations in future, looking into a
198
00:20:06,649 --> 00:20:10,635
person, what where they've traveled and so
on and so forth. But that has nothing to
199
00:20:10,635 --> 00:20:15,980
do with the original purpose of PNR, the
PNR directive. And what at least everybody
200
00:20:15,980 --> 00:20:21,243
here will know in all data storing, so
data storing is in itself a problem. It's
201
00:20:21,243 --> 00:20:25,169
in itself a violation of fundamental
rights when there is no legitimate reason
202
00:20:25,169 --> 00:20:30,316
to do so. But also all data storage puts
the data stored at risk. And as we've
203
00:20:30,316 --> 00:20:34,980
mentioned already, there's the payment
data, especially there's other other
204
00:20:34,980 --> 00:20:39,320
sensitive data with whom you've traveled,
whether you've traveled with light luggage
205
00:20:39,320 --> 00:20:45,787
or not, where you have gone to, via which
place and so on and so forth. Another
206
00:20:45,787 --> 00:20:49,370
point, which is a bit more complicated is
that the director does not sufficiently
207
00:20:49,370 --> 00:20:54,039
differentiate between crimes where
automatic profiling could make sense and
208
00:20:54,039 --> 00:20:59,901
others. So as I have said, there may be a
point in saying that the typical
209
00:20:59,901 --> 00:21:06,039
terrorists would fly from A to B via C
without checking in luggage using this or
210
00:21:06,039 --> 00:21:11,220
that tourist office and so on and so
forth. So maybe just assume that this is
211
00:21:11,220 --> 00:21:17,200
the case. This, no one can can tell me
that there is a typical flight pattern of
212
00:21:17,200 --> 00:21:23,200
a fraudster where you could ask someone
define which way a fraudster typically
213
00:21:23,200 --> 00:21:27,990
flies and identify such a person. So what
the directive would have needed to do if
214
00:21:27,990 --> 00:21:32,269
they wanted had wanted to check against
predetermined criteria would have been to
215
00:21:32,269 --> 00:21:38,943
identify for which crimes - exactly, and
only for these - you can use such a voodoo
216
00:21:38,943 --> 00:21:44,389
miracle weapon. And finally, these are not
the only arguments, but the more most
217
00:21:44,389 --> 00:21:49,269
important ones. We expect that the false
positives especially will lead to
218
00:21:49,269 --> 00:21:55,249
discrimination against minorities. And one
example that the German National Police,
219
00:21:55,249 --> 00:22:01,249
the Bundeskriminalamt has given us for a
predetermined criteria are young men
220
00:22:01,249 --> 00:22:06,732
flying from airports from the south of
Turkey to a major European city. So
221
00:22:06,732 --> 00:22:10,700
they're thinking about former IS fighters,
IS terrorists. And as you can easily
222
00:22:10,700 --> 00:22:15,690
imagine what kind of people will be
sitting in in in a on a plane that's
223
00:22:15,690 --> 00:22:20,310
coming from the south of Turkey to Germany
or to any other European country. Of
224
00:22:20,310 --> 00:22:25,986
course, this will affect them
disproportionately, affect minorities. And
225
00:22:25,986 --> 00:22:32,030
it is already now highly intransparent
what how these these predetermined
226
00:22:32,030 --> 00:22:38,220
criteria are developed. And imagine a near
future where law enforcement will
227
00:22:38,220 --> 00:22:42,640
naturally try to involve artificial
intelligence and finding patterns in the
228
00:22:42,640 --> 00:22:48,179
raw data of flight movements of PNR data,
of the treasure they're now hoarding with
229
00:22:48,179 --> 00:22:54,559
a five year worth of data. And at the
latest, at that point in time, it will be
230
00:22:54,559 --> 00:23:00,710
impossible for us to understand why a
certain criterion was defined and how how
231
00:23:00,710 --> 00:23:04,372
to challenge it when you're in the
position to be arrested at the airport,
232
00:23:04,372 --> 00:23:10,189
for instance. So what can we do? And
that's where we come in. The two
233
00:23:10,189 --> 00:23:15,919
organizations that we are. We are no
typical advocacy organizations, but we do
234
00:23:15,919 --> 00:23:21,039
strategic litigation. Because
unfortunately no advocacy worked on the
235
00:23:21,039 --> 00:23:26,220
PNR directive. It came into force pretty
much as the, um, as national law
236
00:23:26,220 --> 00:23:34,139
enforcement wanted it to be. And so there
is one instance, one authority at the time
237
00:23:34,139 --> 00:23:39,019
that in Europe, in Germany, in Europe, the
European Union, the courts, which can
238
00:23:39,019 --> 00:23:44,919
which can ideally, um, dismiss of the
reasons of the motivations of law
239
00:23:44,919 --> 00:23:51,340
enforcement to have such a directive
enforced and can try to objectively assess
240
00:23:51,340 --> 00:23:57,181
whether this is actually legal and should
remain in force, stay in force or not. And
241
00:23:57,181 --> 00:24:01,639
we did this through litigation both in
Germany and in Austria, and both are
242
00:24:01,639 --> 00:24:06,490
having the same goal, which is to present
to the European Court of Justice the
243
00:24:06,490 --> 00:24:11,774
question whether the PNR directive and any
national law that is transposing the PNR
244
00:24:11,774 --> 00:24:17,980
directive is in violation of the Charter
of Fundamental Rights. Why do we have to
245
00:24:17,980 --> 00:24:23,470
go? Why is the ECJ important? Because when
you have a national law that directly
246
00:24:23,470 --> 00:24:31,330
transposes a European law, a directive,
then then only the ECJ can declare such a
247
00:24:31,330 --> 00:24:35,419
law void. There is no way for, for
instance, in Germany, the federal
248
00:24:35,419 --> 00:24:39,669
constitutional court, the
Bundesverfassungsgericht, to say that this
249
00:24:39,669 --> 00:24:46,016
law should not be applied any longer. This
question must be presented to the ECJ. So
250
00:24:46,016 --> 00:24:50,940
how could we get to the ECJ? This actually
was a process that took us quite a bit of
251
00:24:50,940 --> 00:24:56,489
time. It's been two years in the making. A
year ago, we launched six different
252
00:24:56,489 --> 00:25:01,270
complaints of six different plaintiffs
that are flying all over Europe, that we
253
00:25:01,270 --> 00:25:05,792
booked flights for them that led them to a
European member states, a European Union
254
00:25:05,792 --> 00:25:10,926
member states and two states outside of
the European Union. And we sent the
255
00:25:10,926 --> 00:25:16,029
complaints to three different courts. The
one, two complaints were directed against
256
00:25:16,029 --> 00:25:20,320
the German national police and went to the
administrative court in Wiesbaden, and
257
00:25:20,320 --> 00:25:24,559
four others were directed against the
airplane airlines. So we tried to
258
00:25:24,559 --> 00:25:30,950
diversify as much as possible in order to
find a judge that would agree with us that
259
00:25:30,950 --> 00:25:36,779
this is problematic and this needs
checking. And we are optimistic that
260
00:25:36,779 --> 00:25:43,500
either the court in Wiesbaden or the court
in Cologne will soon present these very
261
00:25:43,500 --> 00:25:48,289
questions to the court, whether the German
transposition law and the PNR directive
262
00:25:48,289 --> 00:25:53,119
itself are violating fundamental rights
after European of the Charter of the
263
00:25:53,119 --> 00:25:58,981
European Union.
Walter: So as Bijan already mentioned, our
264
00:25:58,981 --> 00:26:05,389
aim is to bring our case as quick as
possible to the European Court of Justice.
265
00:26:05,389 --> 00:26:11,470
So we had different options. And in Austria,
we went a third way. We brought a case
266
00:26:11,470 --> 00:26:18,976
before the Austrian Data Protection Authority
against the Fluggastdatenzentralstelle
267
00:26:18,976 --> 00:26:24,603
im Bundeskriminalamt, a passenger
named unit. And we we brought several
268
00:26:24,603 --> 00:26:31,059
different cases and we also found out that
different, smaller things which we are on.
269
00:26:31,059 --> 00:26:38,014
But the main thing is that this case
already went as planned to the
270
00:26:38,014 --> 00:26:46,579
Bundesverwaltungsgericht, so the federal
administrative court in Austria. And from
271
00:26:46,579 --> 00:26:54,850
there, we hope that is also soon forwarded
to the European Court of Justice. And
272
00:26:54,850 --> 00:27:02,239
theoretically, it would be enough if one
case hits the European Court of Justice.
273
00:27:02,239 --> 00:27:07,590
But practically, it is, of course, very
important to have different strategies
274
00:27:07,590 --> 00:27:15,749
because there are different speeds and so
on. So that's why we also should mention
275
00:27:15,749 --> 00:27:22,929
another case, the the Belgian case. So
this Belgian human rights organization,
276
00:27:22,929 --> 00:27:28,647
they also brought the case before a
Belgian court. In this case, it was
277
00:27:28,647 --> 00:27:34,720
directly the Belgian constitutional court.
So they had a direct way to the
278
00:27:34,720 --> 00:27:40,340
constitutional court, unlike our cases in
Austria, where this or in Germany where
279
00:27:40,340 --> 00:27:47,369
this was not possible. And therefore, the
Belgian constitutional court already
280
00:27:47,369 --> 00:27:55,037
referred this case to the European Court
of Justice. And we are hoping that our
281
00:27:55,037 --> 00:28:01,210
case will be soon or cases, or at least
some of them will soon be joined with this
282
00:28:01,210 --> 00:28:11,220
case at the European Court of Justice, and
then decided together. So to sum up, we
283
00:28:11,220 --> 00:28:20,429
have actually a very infringing piece of
legislation the PNR directive, PNR
284
00:28:20,429 --> 00:28:28,529
processing, as Bijan explained to us in
more detail, is extremely intrusive in all
285
00:28:28,529 --> 00:28:34,549
flight passengers' fundamental rights. It
violates fundamental rights, especially
286
00:28:34,549 --> 00:28:41,460
because it is already... is also
ineffective and disproportionate. So we
287
00:28:41,460 --> 00:28:47,649
heard about these different things. The
base rate fallacy that it is ineffective
288
00:28:47,649 --> 00:28:54,099
and disproportionate because it is not
really possible to find specific suspects
289
00:28:54,099 --> 00:29:02,570
in such amount of data with without having
a lot, a real lot of false positives. So
290
00:29:02,570 --> 00:29:08,190
other arguments are that it is data
retention in the first place. So also
291
00:29:08,190 --> 00:29:14,989
already the retention of the data of
people like you and me is a big problem
292
00:29:14,989 --> 00:29:22,600
and unlawful. And this general suspicion
it leads to. So everybody becomes a
293
00:29:22,600 --> 00:29:30,200
suspect and can become practically a
suspect, can get problems practically from
294
00:29:30,200 --> 00:29:38,899
that legislation without being a criminal.
And yeah, we have strong arguments as we
295
00:29:38,899 --> 00:29:48,599
showed you already, the case of the Canada
PNR directive, the PNR agreement with
296
00:29:48,599 --> 00:29:55,259
Canada is very similar in practice to the
PNR directive. So the arguments already
297
00:29:55,259 --> 00:30:01,480
held before the European Court of Justice.
So actually, it's a shame that this was
298
00:30:01,480 --> 00:30:07,971
not stopped earlier. And civil rights
organizations as we are have to do that.
299
00:30:07,971 --> 00:30:16,590
And that's what we do. And that's also why
we depend on donations. So that's also
300
00:30:16,590 --> 00:30:21,789
important to stress that our work people
having people fully employed to do things
301
00:30:21,789 --> 00:30:28,700
like that cost some money. And that's
where you can find us. So we have a
302
00:30:28,700 --> 00:30:35,799
campaign website, nopnr.eu in German
and English. And you can find us, of
303
00:30:35,799 --> 00:30:40,580
course, on our website and both websites
and find ways how to join us, how to
304
00:30:40,580 --> 00:30:46,730
support us. And also still today, you can
meet us at our assembly in the CCL
305
00:30:46,730 --> 00:30:52,590
building the about freedom assembly, where
both the Gesellschaft für Freiheitsrechte
306
00:30:52,590 --> 00:31:00,659
and Epicentre Works have their desk and
you can ask all the question. But first,
307
00:31:00,659 --> 00:31:03,378
ask all your questions now. Thank you.
308
00:31:03,378 --> 00:31:06,870
Applause
309
00:31:06,870 --> 00:31:14,760
Herald: Thank you, Walter and Bijan, for
this very clarifying statements. I suppose
310
00:31:14,760 --> 00:31:19,379
there are quite some questions here in the
audience. Only I'm looking at someone
311
00:31:19,379 --> 00:31:32,252
who's grabbing a microphone now. I see the
signal angel. Yes. The mic is not on. Can
312
00:31:32,252 --> 00:31:47,591
someone help him? Signal Angel needs a mic.
Yes, it's almost there. Brains are
313
00:31:47,591 --> 00:31:51,409
working.
Signal Angel: Thank you. Is there a cheap
314
00:31:51,409 --> 00:31:55,470
method to spam for some trees, for
example, by booking flight under a false
315
00:31:55,470 --> 00:32:02,999
name and then canceling the flight?
Bijan: Well, I think it's it's difficult
316
00:32:02,999 --> 00:32:06,889
to say. I didn't get the very first words.
Sorry.
317
00:32:06,889 --> 00:32:11,159
Signal Angel: Yes, the very first one was:
is there a cheap method to spam, to spam
318
00:32:11,159 --> 00:32:14,809
for some trees?
Bijan: Yeah. Theoretically, I don't think
319
00:32:14,809 --> 00:32:19,129
that anything could speak against that.
Yeah, but the problem is that you would
320
00:32:19,129 --> 00:32:23,749
need to cancel very late because, um, I
think the first time they push the data,
321
00:32:23,749 --> 00:32:28,139
the airlines are pushing the data to the
national police is, 48 hours before the
322
00:32:28,139 --> 00:32:32,799
before boarding. So that might come to
become a bit expensive.
323
00:32:32,799 --> 00:32:34,439
Laughter
324
00:32:34,439 --> 00:32:38,918
Walter: I would want to make a general
remark also on that. Of course, here,
325
00:32:38,918 --> 00:32:44,109
especially here, thoughts like that, how
to hack the system are very important and
326
00:32:44,109 --> 00:32:50,759
can help. But our general approach is to
take legal action to protect all people at
327
00:32:50,759 --> 00:32:56,019
the same way, and not only those who who
are able to protect themselves or hack the
328
00:32:56,019 --> 00:33:04,139
system or whatever. So that's the reason
why we both go this general way to bring
329
00:33:04,139 --> 00:33:11,449
that down. Completely.
Herald: And other question here. Yes.
330
00:33:11,449 --> 00:33:20,879
Sorry, sir. Please.
Q: What do you expect as a result of your
331
00:33:20,879 --> 00:33:26,549
litigation if you are successful in court?
Will ... do you expect the courts to
332
00:33:26,549 --> 00:33:34,309
strike down the directive entirely, or do
you expect another legislative process to
333
00:33:34,309 --> 00:33:42,260
do the same thing again or to fix, quote
unquote, the directive in very small ways
334
00:33:42,260 --> 00:33:47,570
just to to drag out this battle and
continue the practice. What do you think
335
00:33:47,570 --> 00:33:52,210
the effects will be?
Bijan: Well, we think that the European
336
00:33:52,210 --> 00:33:56,070
Court of Justice, if it follows our
argument, our reasoning, it should it will
337
00:33:56,070 --> 00:33:59,690
strike down the PNR directive entirely,
because the way it is set up is
338
00:33:59,690 --> 00:34:06,561
fundamentally not in in accordance with
what it earlier ruled so far. Unless it
339
00:34:06,561 --> 00:34:10,550
will change its its entire history of
ruling on data retention and so on and so
340
00:34:10,550 --> 00:34:15,679
forth. But of course, we will expect the
member states to push for another
341
00:34:15,679 --> 00:34:21,349
legislation that may be similar, but not
the exact same thing. So I can imagine
342
00:34:21,349 --> 00:34:25,889
something of a of the sort of data
retention of telecommunications, as it
343
00:34:25,889 --> 00:34:31,060
were, and with airlines retaining the data
and keeping it for a shorter period of
344
00:34:31,060 --> 00:34:36,291
time and only giving it out when there is
a specific request with, where there is a
345
00:34:36,291 --> 00:34:41,029
specific reason for law enforcement to ask
for the data. I could imagine such a thing
346
00:34:41,029 --> 00:34:45,845
coming up again and then we would need to
check whether this is illegal or not. And
347
00:34:45,845 --> 00:34:50,254
maybe go through the whole procedure as
well. But it is it would be an immense
348
00:34:50,254 --> 00:34:56,191
success if the PNR directive as it stands
would be void. Declared void.
349
00:34:56,191 --> 00:35:02,290
Herald: Thank you. Someone else has a
question. I see the person here.
350
00:35:02,290 --> 00:35:10,789
Microphone one, please.
Q: Hel-lo, yeah. Okay, so you had the
351
00:35:10,789 --> 00:35:15,480
agreement that, uh, there are a lot of
false positives when they checked up PNR
352
00:35:15,480 --> 00:35:20,640
data. Um, do we have any information how
long it takes for them to react on the PNR
353
00:35:20,640 --> 00:35:25,570
data if they get a positive hit? So maybe
they won't react after the person has
354
00:35:25,570 --> 00:35:31,109
landed and already, uh, is in the country?
Bijan: They claim that they can act
355
00:35:31,109 --> 00:35:36,930
immediately, but we can't know that for
sure. So the fact that they had 57 arrests
356
00:35:36,930 --> 00:35:42,366
at the airports signals that at least that
in some respects this is true. But we
357
00:35:42,366 --> 00:35:47,477
cannot know for sure how much, how quickly
they they they kind of react. And keep in
358
00:35:47,477 --> 00:35:52,470
mind, this is only the start. So, so far
in Germany, right up until the point where
359
00:35:52,470 --> 00:35:57,220
this the data that I presented for Germany
came about, there were only 9 airlines, I
360
00:35:57,220 --> 00:36:01,440
think, that were linked to the system. So
expect there to be much more data coming
361
00:36:01,440 --> 00:36:05,829
in. And once they start with a
predetermined criteria thing, this will
362
00:36:05,829 --> 00:36:13,440
multiply probably. Um, even so, I cannot
imagine unless they they ... have this
363
00:36:13,440 --> 00:36:20,279
new, um, thing with hundreds of people
involved that they can act immediately in
364
00:36:20,279 --> 00:36:26,240
each and every case.
Herald: Thank you. There is a question
365
00:36:26,240 --> 00:36:30,100
again on the Internet. Yes.
Signal Angel: Yes. How come, you haven't
366
00:36:30,100 --> 00:36:35,349
tried voiding the local at one provisions
that this PNR there for intra EU flights? (???)
367
00:36:35,349 --> 00:36:39,180
That seems most likely against Schengen
provisions.
368
00:36:39,180 --> 00:36:44,837
Bijan: We have addressed that as well. We
have picked intra-EU flights also. We have
369
00:36:44,837 --> 00:36:52,290
not just picked flights that go extra EU,
but, we've also made the point about the
370
00:36:52,290 --> 00:36:57,839
the violation of Schengen criteria. But
that is not so much that is not the focus
371
00:36:57,839 --> 00:37:02,893
of our argument because they are, in our
opinion, much stronger ones. Because with
372
00:37:02,893 --> 00:37:06,990
Schengen you would need to argue that it's
practically impossible to enter the
373
00:37:06,990 --> 00:37:13,260
country without being held up and you're
not being held up in a physical form, at
374
00:37:13,260 --> 00:37:18,910
least not in general, generally. And so
this argument is a bit more difficult than
375
00:37:18,910 --> 00:37:25,024
having an actual border checking of
people. But we're making this point, of
376
00:37:25,024 --> 00:37:30,460
course. And but we rely on other points
that we think are stronger.
377
00:37:30,460 --> 00:37:35,540
Herald: Okay. Please. Microphone number
one, please.
378
00:37:35,540 --> 00:37:40,093
Q: Is there also data being collected on
flights inside a country. So, for example,
379
00:37:40,093 --> 00:37:43,530
from Munich to Berlin.
Bijan: Not yet. Not under the directive.
380
00:37:43,530 --> 00:37:48,099
And theoretically, of course, that the
German legislator or any other legislator
381
00:37:48,099 --> 00:37:51,820
could decide to include that as well, but
not so far.
382
00:37:51,820 --> 00:37:56,329
Herald: Number two, please. Microphone.
Yeah.
383
00:37:56,329 --> 00:38:01,920
Q: I was wondering how much, uh, false
negatives are in there. You know, that,
384
00:38:01,920 --> 00:38:07,940
like, uh, these big databases. If I don't
act like a normal terrorist or something
385
00:38:07,940 --> 00:38:10,570
than I am?
Bijan: We don't we don't know,
386
00:38:10,570 --> 00:38:15,130
unfortunately, not yet. Um, I did. I think
it would be very interesting, especially
387
00:38:15,130 --> 00:38:21,750
for the predetermined criteria , to see
how many they miss. Um, but yeah. No, not
388
00:38:21,750 --> 00:38:28,579
nothing at.
Herald: Yeah, and there is no undo button,
389
00:38:28,579 --> 00:38:33,579
I think. No. No. No undo. That's always
the thing that I that I'm worried about,
390
00:38:33,579 --> 00:38:38,140
you know. Then you have an announcement
about France's data that go out and then
391
00:38:38,140 --> 00:38:44,040
you can't have an undo. So what do we do
then? It's always new. Yeah, you can keep
392
00:38:44,040 --> 00:38:48,640
this for five years now. But who says it's
there for five years and what kind of
393
00:38:48,640 --> 00:38:51,967
interpretation to get out of it for five
years? After five years?
394
00:38:51,967 --> 00:38:55,089
Bijan: You can't know in
which database you will be transferred in
395
00:38:55,089 --> 00:38:59,940
the meantime, because law enforcement can
access the data of that very data set and
396
00:38:59,940 --> 00:39:03,700
forth for that data and the PNR data set
and put it in another data set because
397
00:39:03,700 --> 00:39:08,480
they have whatever reason to do so. And
then these are again enlarged and
398
00:39:08,480 --> 00:39:12,670
enlarged. And then you will find another
reason why they should remain in there for
399
00:39:12,670 --> 00:39:17,230
a longer time. So, yeah. That's why we're
fighting this now and hoping to change the
400
00:39:17,230 --> 00:39:19,480
future.
Herald: How do you see your chances?
401
00:39:19,480 --> 00:39:27,070
Actually, uh, a long term or short term
chances to get to that point is that?
402
00:39:27,070 --> 00:39:30,500
Bijan: We are very convinced that we will
be successful, because otherwise we
403
00:39:30,500 --> 00:39:33,529
wouldn't have started this. This is one of
our principles. We only do things that we
404
00:39:33,529 --> 00:39:38,321
are convinced of being able to win and we
think that we will win this. And what will
405
00:39:38,321 --> 00:39:42,240
come out of it? Referring to the I think
the second and the second question
406
00:39:42,240 --> 00:39:47,480
earlier. And what will be happening in the
future with other legislation? I can't
407
00:39:47,480 --> 00:39:51,430
know. But one argument the police is
always making or in private, at least to
408
00:39:51,430 --> 00:39:55,819
me, are is that they're saying, well,
people will get used to it and it won't be
409
00:39:55,819 --> 00:40:00,783
in in five or 10 years. Nobody's gonna be
wondering about things like this. And this
410
00:40:00,783 --> 00:40:05,299
is exactly what we are working against,
that this never becomes normal, because if
411
00:40:05,299 --> 00:40:07,470
this becomes normal, as I've argued
before,
412
00:40:07,470 --> 00:40:11,130
applause
Herald: needs an applause Yes.
413
00:40:11,130 --> 00:40:14,880
Bijan: If it becomes normal, as I've
argued before, it is easy to extend it to
414
00:40:14,880 --> 00:40:20,529
all sorts of life and ways of life and
walks of life. And this then would be in a
415
00:40:20,529 --> 00:40:25,859
surveillance state par excellence.
Herald: We were very close there. So we
416
00:40:25,859 --> 00:40:30,089
need to support them really hard. There is
one last question I suggest. No. There is
417
00:40:30,089 --> 00:40:35,619
two questions. Number two. Yes.
Q: Does the PNR directive apply only for
418
00:40:35,619 --> 00:40:39,492
regular scheduled flights? So does it also
apply for private flights? The general
419
00:40:39,492 --> 00:40:42,819
aviation business flights, etc.?
Bijan: Good question. I don't know.
420
00:40:42,819 --> 00:40:48,510
Actually, I look into that and. Write me!
Come, come here later and I'll check and
421
00:40:48,510 --> 00:40:52,310
I'll give you an answer.
Herald: Then there is one at number one.
422
00:40:52,310 --> 00:40:56,710
Q: I just wanted to ask a question in
response to the idea that this is becoming
423
00:40:56,710 --> 00:41:01,810
very normal, because one thing that I
think has become very normal that hasn't
424
00:41:01,810 --> 00:41:07,924
been mentioned explicitly is the idea that
people can be essentially put on a watch
425
00:41:07,924 --> 00:41:12,782
list as being a potential criminal in the
absence of a crime. And we have these
426
00:41:12,782 --> 00:41:19,605
terrorist watch lists all over the world
now. That is now the new normal. And I
427
00:41:19,605 --> 00:41:24,750
think that's very problematic. And can you
just maybe talk about: Do we, do you see a
428
00:41:24,750 --> 00:41:30,829
future where we can actually get back to,
you know, only arresting or investigating
429
00:41:30,829 --> 00:41:33,890
people because of probable cause, for
example?
430
00:41:33,890 --> 00:41:39,421
Bijan: Oh, I hope that this will be our
future. But, uh, about that point, that
431
00:41:39,421 --> 00:41:43,750
very point, I'm not too optimistic, to be
honest. I am optimistic about one other
432
00:41:43,750 --> 00:41:48,859
one. Another thing that is that these
instruments that are now being created
433
00:41:48,859 --> 00:41:52,910
will prove to be highly ineffective, as
we've so now see now already with checking
434
00:41:52,910 --> 00:41:58,290
against databases, that is already a lot
of work and very tedious work. But with
435
00:41:58,290 --> 00:42:04,079
the idea that you can define criteria for
people that that are legitimately to be
436
00:42:04,079 --> 00:42:08,710
suspected of committing a crime in the
future, I think it will prove, at least
437
00:42:08,710 --> 00:42:12,990
for the next few decades, to be quite
impossible. And this is I don't know if
438
00:42:12,990 --> 00:42:19,880
this came across correctly sufficiently,
but this is really the core issue that we
439
00:42:19,880 --> 00:42:25,619
have with the PNR directive. They are
claiming that they can find suspects of
440
00:42:25,619 --> 00:42:31,690
crimes or future crimes. Imagine! Not not
someone that has committed a crime or that
441
00:42:31,690 --> 00:42:36,890
will definitely commit a crime, but that
can reasonably be suspected of committing
442
00:42:36,890 --> 00:42:43,960
a crime in the future, and then act upon
that. And that is really a huge step into
443
00:42:43,960 --> 00:42:49,220
what I called voodoo, about the
expectation that you can take data and
444
00:42:49,220 --> 00:42:55,400
prevent crime. Minority Report times.
Yeah. To the power five. I don't know.
445
00:42:55,400 --> 00:43:00,534
Herald: Sit back and relax. Thank you
Bijan and thank you, Walter, for this
446
00:43:00,534 --> 00:43:08,520
fantastic lecture. Please support them at
noPNR dot EU, go to their booth as well.
447
00:43:08,520 --> 00:43:10,520
And thank you all.
448
00:43:10,520 --> 00:43:22,010
36C3 postroll music
449
00:43:22,010 --> 00:43:38,000
Subtitles created by c3subtitles.de
in the year 2021. Join, and help us!